PIN Detection And Plug-in Positioning Of Complex Dense Pin Components Based On Machine Vision

Along with the industrial revolutions,China's manufacturing industry has developed rapidly.As a basic industry in the manufacture of electronic information products,the PCB industry has complicated and diversified its internal components based on the substantial increase in demand and output.Due to the regular shape of general standard parts,their detection and assembly techniques are relatively mature.However,due to factors such as diversified shapes,special-shaped parts have caused corresponding difficulties in the process of positioning,testing and insertion.Specifically,the development of complex dense multi-pin shaped components in the field of industrial inspection has been slow.These components will have defects such as missing pins,skew,bend,and height inconsistencies in the manufacturing process,affecting subsequent plug-in,assembly and other processes.The existing pin positioning and detection mainly have the following problems:1.Due to the characteristics of slender pins and small cross-sectional area,traditional image acquisition makes the bottom reflection serious,and high-quality pin images cannot be obtained;2.When the number of pins increases,the arrangement is dense At the same time,the problems such as overlapping of the pins increase the positioning error;3.For the lack of pins,skew,bending and other different defects can't be solved at one time,the lack of a unified detection algorithm.In order to solve the above problems,this article aims to study a set of methods based on machine vision for the detection of complex dense pin components and plug-in positioning.The main research work is as follows:On the basis of fully analyzing the characteristics of complex and dense pin components,the design of 2D and 3D vision systems was carried out,and the final collection plan was clarified by combining the collected pin images.The pin data is obtained by three-dimensional laser scanning,and the data is masked with a threshold in the height direction,and the generated three-dimensional point cloud data is converted into a two-dimensional pin image.Finally,a high-quality complex dense pin map is obtained.A positioning algorithm for complex dense pin components is proposed.Mathematical description and model establishment are performed for each positioning module,including the mathematical model of the pin component retrieving and the PCB generated when the board enters the board.The offset is solved and the angle and position are corrected.Through the system calibration,the coordinate systems are unified,the camera calibration is introduced during the calibration process,and the hand-eye calibration method of the robot arm is derived,and the mapping relationship between the image coordinates and the robot arm coordinates is finally obtained.A detection algorithm for complex and dense pin components is proposed,and the collected pin images are processed through contour retrieval and other corresponding image processing methods to generate the required detection map.The detection map and the template map are subjected to pixel-by-pixel sum calculation after image registration transformation,and then,the purpose of effectively detecting pin defects is achieved by counting the number of pixels.Finally,by constructing a complete picking,testing and pin component insertion system,the accuracy and system accuracy of the method in this paper are verified.Experiments show that the method in this paper can solve many defects of complex dense pin components at once.When the number increases,it can still be effectively positioned and tested to meet the needs of industrial testing.